|
Abstract:
|
We introduce a novel method for modeling space-time infectious disease count data with covariates, building on the discrete time endemic-epidemic model proposed by Held et al. (2005). The model contains three components with infection for a susceptible occurring from within their own area, a neighboring area, or the environment. The current approach includes random effects that may be added, for each discrete region, to any or all of the three components. Our method allows the underlying disease map to be modeled in continuous space using a stochastic partial differential equations approach. By modeling the spatial surface as continuous and incorporating covariate information, we can gain an understanding of how the space-time epidemic evolves. We illustrate the method using a German influenza dataset.
|
